Jue Zhang of Griffith University

Weakly supervised object segmentation such as photovoltaic panel mapping in remote sensing suffers from noisy pseudo labels that erode feature discrimination and reduce accuracy. This talk presents three complementary advances to counteract label noise and strengthen generalization. First, we introduce a self paced residual aggregation network that explicitly learns under pseudo label noise, progressively down weighting unreliable signals to improve segmentation accuracy. Second, we adopt a principled and statistically grounded decomposition that treats pseudo labels as clean latent labels passed through a noise process, enabling interpretable noise modeling, bias correction, and approximate clean label recovery. Third, we design a dual stream decoder that fuses heterogeneous pseudo labels for collaborative supervision, yielding boundary aware discrimination and complementary joint learning. Together, these components deliver consistent gains in weakly supervised settings and translate to practical benefits for energy GIS, renewable energy siting optimization, and ecological and environmental assessment. This research program has produced three journal papers including one in IEEE Transactions on Image Processing and two in IEEE Transactions on Geoscience and Remote Sensing, as well as four conference papers.

Biography:

Dr Jue Zhang Member IEEE received the Bachelor of Science degree in Electronic Science and Technology from Beijing Normal University in 2014 and the Doctor of Philosophy degree in Computer Science from the University of New South Wales Canberra in 2024. She is currently a Research Fellow with the ARC Research Hub for Driving Farming Productivity and Disease Prevention at Griffith University Australia. Her research interests include remote sensing machine learning and weakly supervised learning. Dr Zhang received the Jose Bioucas Dias Outstanding Paper Award at WHISPERS 2023. She has served as Chapter Chair of the IEEE ACT Section UNSW Student Branch and the IEEE GRSS Student Chapter and is currently the Young Professionals Representative for IEEE GRSS in Australia and New Zealand and a member of the IGARSS 2026 Local Organizing Committee. She also served as Assistant to the Editor in Chief for IEEE Transactions on Geoscience and Remote Sensing.

Email: